In a 15-month study, mature-length foliage samples were collected at 2-3 week intervals from the lower branch clusters of cuttings of nine non-select Pinus radiata
clones. The cuttings which had been planted out four years previously were growing on a uniform pumice soil site at Rotorua.
Foliage samples representative of each clone on 25 sampling dates were analysed for 10 nutrient elements. Analysis of variance of the foliar data indicated that, despite the confounding for ramets within clones and dates, the effects of date and clone respectively were in almost all instances very highly significant (p < .001). In order of magnitude, the fractions of total variance accounted for by clone were: .48 (B), .37 (K, Zn), .23 (Ca, Mg), .12 (Mn), .07 (Na), .09 (N), .04 (P) and .02 (Cu); those accounted for by sampling date were: .77 (Cu), .68 (N), .60 (P), .54 (Na), .41 (Ca), .37 (Mg), .29 (Mn), .22 (K, B), and .11 (Zn).
Clonal repeatabilities as at any given date varied greatly with element, being highest for B and K (.60, .46), intermediate for Zn, Ca and Mg (.39, .37 and .35), low for N and Mn (.26, .17) and very low for Na, Cu and P (.11, .06 and .05). Clone-season interaction appeared to be minor. Tree-to-tree coefficients of variation also varied widely with element and were (in order of magnitude): .32 (Na), .30 (Mn), .26 (B), .25 (Zn), .22 (Ca), .20 (Mg), .14 (K), .12 (P, Cu), .09 (N). To detect differences (p < 0.05) of at least 20% between the means of two sample populations for macronutrients would evidently require sampling from at least 10 trees per composite population; for micronutrients the number generally seems higher (c. 15-20).